Freenet is not completely untraceable, it's just really hard to tell who originally published a file to it (but not so hard to tell who republishes files, although it is hard to prove if they intended to republish or not), and really hard to prove someone requested something from it (although trivial for their neighboring nodes to tell that they got something from it, proof that they requested to get it is the more difficult part to prove). There are attacks on Freenet though. It is also much less studied than Tor is. I think it is probably a lot better in some ways and a lot worse in others, but it has a ton more 'unknowns' than Tor does, at this point. It still aims to achieve anonymity, but it does so in an almost entirely different way than Tor does. Freenet aims to make it impossible to prove who originally published a file, and attempts to make it impossible to prove who intentionally republishes files and who intentionally requests to download files (ie: Freenet aims primarily to maintain plausible deniability, 'yes I downloaded that, but can you prove that I requested it?' 'Yes, I sent that from my computer and stored it on my computer, but can you prove that I did so intentionally?', 'Yes, I sent that file, but can you prove I was the one who originally added it to the network?' ). Freenet also aims to be highly censorship resistant by storing content distributed over many servers through out the network (ie: 'you can't censor this site because even though you can identify the nodes that store it easily enough, there are hundreds of them around the world, and you can't determine if the people running the nodes are aware of what they are storing, and even after you take a few of them down the content just shifts to others) where as with Tor it is more like 'You don't know what I downloaded because it is layer encrypted on the path back to me' (but if you follow it back to me with a timing attack, or identify it with a traffic classifier at my entry node, you can essentially prove that I intended to download it) ... 'You don't know if I uploaded this file or not because you can't trace me back through the network' (but if you manage to trace me back you can essentially prove that I uploaded the file) .... 'You can't determine who I communicate with because our layer encrypted communications are routed through lots of nodes' (but if you do manage to get in a position to do a linking attack you can confirm who is talking with who) .... 'You can't censor this website because you (allegedly) can't easily find the server to take it down' (but if you do trace the server you can take it down) Attacks against Tor tend to try to observe a packet at multiple points as it travels through the network, in an attempt to link the sender of the packet to the receiver or vice versa (or to identify the content of the sent / received packet). Attacks against Freenet tend to use statistical formulas to come to an estimate regarding the probability of a certain node requesting or inserting a certain file or type of content (after all, it is trivial for all neighboring Freenet nodes to determine the content that they relay on to and from their neighbors, it is just hard for them to determine if the neighbor inserted/requested the content in the first place....) The only place Freenet is a clear winner over Tor is in censorship resistance....good luck removing popular content from Freenet.